A novel combined regularization algorithm of total variation and Tikhonov regularization for open electrical...

Area of Science:

• Medical Imaging
• Computational Electromagnetics

Background:

• Tikhonov regularization in electrical impedance tomography (EIT) yields smooth reconstructions, hindering clear separation of internal structures.
• Total variation (TV) methods, like lagged diffusivity (LD), offer edge sharpening and noise robustness but have limited convergence regions.

Purpose of the Study:

• To develop and evaluate a novel regularization method combining Tikhonov and LD for improved EIT image reconstruction.
• To investigate the impact of a weighted parameter on Tikhonov regularization for varying inclusion depths.
• To assess algorithm performance with noisy data and analyze current injection patterns.

Main Methods:

• Implementation of Tikhonov, LD, and a combined Tikhonov-LD regularization method for 2D open EIT.
• Introduction of a weighted parameter in Tikhonov regularization to analyze its effect on image resolution and quality.
• Performance evaluation using noisy datasets and analysis of different current injection patterns.

Main Results:

• The combined Tikhonov-LD regularization method demonstrates stable convergence, superior reconstruction quality, sharper contrasts, and enhanced noise robustness over individual methods.
• A weighted parameter in Tikhonov regularization affects image resolution and quality based on inclusion depth.
• Modified current injection patterns, particularly those with larger driver angles, improve reconstruction quality.

Conclusions:

• The combined Tikhonov-LD regularization method significantly advances EIT image reconstruction accuracy and robustness.
• Optimizing current injection patterns is crucial for enhancing EIT imaging performance.
• This combined approach offers a promising solution for clearer and more reliable EIT imaging in various applications.